Researchers from Paderborn University have discovered that dangling bonds in silicon, considered defects, can increase the efficiency of solar cells by up to 40% when combined with tetracene. The finding, published in Physical Review Letters, could raise efficiency from 22% to 31%.
A team from the University of Paderborn (Germany) used the Hawk supercomputer at the HLRS centre in Stuttgart to simulate the interface between silicon and a layer of tetracene, an organic semiconductor. The results, published in Physical Review Letters, show that the silicon defects that the industry has been trying to eliminate for decades can actually facilitate the capture of more light energy, as reported by ecosistemastartup.com.
Dangling bonds are silicon atoms at the interfaces that are not fully bonded to their neighbours. In the classic design of solar cells, they are treated as defects to be eliminated because they act as traps for the electrons generated by light. The simulation by the team led by Marvin Krenz and coordinated by Wolf Gero Schmidt showed the opposite: at the interface between silicon and tetracene, these dangling bonds act as the bridge that allows energy to be transferred from the organic layer to the silicon.
The mechanism is singlet fission. When a high-energy photon strikes the tetracene, the excess energy (about 3 eV above the 1.1 eV that silicon can convert) is normally lost as heat. Singlet fission allows that photon to generate two low-energy excited states (triplets) that are transferred to the silicon, producing two electrons instead of one.
An average solar panel today converts about 22% of the incident energy into electricity. Schmidt estimates that consistently applying singlet fission could multiply efficiency by 1.4, bringing hybrid tetracene-silicon cells close to 31%. To put it in perspective: solar efficiency has improved by about 1% per year over the last two decades, according to the analysis of the finding itself. A nine-point jump is not an incremental improvement; it is a scale change.
This is not an isolated effort. In October 2025, researchers from the University of New South Wales (UNSW) published in ACS Energy Letters an advance with a stable molecule called DPND that, according to pv magazine, could raise the efficiency of silicon modules above 30% and, in their projections, up to 42%. The same work noted that tetracene is photochemically unstable and therefore not suitable for commercial applications, a fact to keep in mind when reading the Paderborn finding.
In March 2026, a team from Kyushu University and Johannes Gutenberg University of Mainz reported in the Journal of the American Chemical Society a molybdenum metal complex capable of capturing energy multiplied by singlet fission with quantum yields of around 130%, according to ScienceDaily.
And in the field of complementary materials, EPFL and CSEM certified in March 2026 a triple-junction perovskite-silicon cell with 30.02% efficiency, surpassing the previous record of 27.1%, according to WWWhat’s New. Conventional silicon panels range between 20% and 24%, and III-V space cells reach 37% but cost about 1,000 times more per watt.
The gap between laboratory and industrial production is the usual obstacle, and here there are two specific challenges. First, tetracene is an organic material that degrades faster than silicon under prolonged exposure to sunlight. Second, depositing a homogeneous layer of tetracene on silicon at an industrial scale without introducing other defects that would reduce efficiency is a manufacturing engineering problem that remains unresolved.
Silicon dominates about 95% of the solar market. That is why the Paderborn finding matters: it suggests that the path to improvement is not only in new materials like perovskite but also in better exploiting the installed base.
For entrepreneurs in the energy sector, first-principles simulation is solidifying as a design tool. The case of Hawk demonstrates that simulating the quantum behaviour of an interface at device scale requires millions of calculations for every picosecond of evolution.
You may also like
